Quantitative evaluation of factors influencing the acquisition of 3D obstetric US images through an analysis of 3D fetal facial scans
Ashanda Connor, King's College London
Objective of study
Three-dimensional (3D) prenatal ultrasound scans of a baby’s facial features are increasingly popular in private practice and there is an emerging role in clinical practice. However, prospective parents are often disappointed with the quality of images obtained. Companies tend to advertise using unrepresentative "best quality" images, leading to a mismatch between the customers’ expectations and the achievable quality of the scan due to difficulties for clinically acquired images. The aim is to explore the impact of maternal-fetal factors on image quality and evaluate scanning strategies that result in optimal 3D facial imaging.
Methods
This retrospective image review study analyses an existing database of 3D surface-rendered ultrasound images acquired from pregnant women as souvenir pictures during a research study. The overall clarity of each image for facial visualisation was scored from 0 (unidentifiable) to 3 (good), and rater agreement between experienced sonographers was assessed. Fifteen maternal and fetal factors as explanatory variables affecting image quality were also scored and statistically assessed.
Results
Images from 41 women (total 342 images, GA range 21–30, BMI 19–37.5 kg/m2) demonstrated only 3% of scans were scored as good. The rating by two obstetric sonographers showed moderate agreement, Cohen’s Kappa = 0.516, p <0.001. Statistically significant factors affecting quality on a Chi-squared test included gestational age Uterine–Fetal Amniotic Assessment (UFAA), placental site, fetal parts obscuring face, maternal age and fetal position (p<0.05). The multiple logistic regression model suggested that these optimal factors in combination also produced the best-quality images (p<0.05) apart from gestational age. The remaining factors did not have a significant impact on quality in this sample.
Conclusion
Good amniotic fluid level around the face, cephalic fetal position, lack of obscuring fetal parts and posterior placenta are positive factors influencing the image quality of 3D obstetric facial views.